From ab39ed11fcc0e56dfa8d05b42e03c0d0417092f6 Mon Sep 17 00:00:00 2001
From: Mengqin Shen <mengqin@google.com>
Date: Thu, 5 Feb 2026 21:56:00 -0800
Subject: [PATCH 1/3] fix(py): fix model config consistencies for gemini

---
 .../genkit/src/genkit/ai/_base_async.py       |  72 +++--------
 .../plugins/google_genai/models/gemini.py     | 116 ++++++++++++------
 .../plugins/google_genai/models/utils.py      |  37 +++++-
 py/samples/google-genai-hello/src/main.py     |   6 +-
 4 files changed, 129 insertions(+), 102 deletions(-)

diff --git a/py/packages/genkit/src/genkit/ai/_base_async.py b/py/packages/genkit/src/genkit/ai/_base_async.py
index 8f849c65b2..c232b57ff8 100644
--- a/py/packages/genkit/src/genkit/ai/_base_async.py
+++ b/py/packages/genkit/src/genkit/ai/_base_async.py
@@ -16,9 +16,8 @@
 
 """Asynchronous server gateway interface implementation for Genkit."""
 
-import signal
 from collections.abc import Coroutine
-from typing import Any, TypeVar
+from typing import Any, TypeVar, cast
 
 import anyio
 import httpx
@@ -127,7 +126,7 @@ async def dev_runner() -> T:
             assert spec is not None
             # Capture spec in local var for nested functions (pyrefly doesn't narrow closures)
             server_spec: ServerSpec = spec
-            user_result: T | None = None
+            user_result: T = None  # type: ignore[assignment]
             user_task_finished_event = anyio.Event()
 
             async def run_user_coro_wrapper() -> None:
@@ -135,64 +134,25 @@ async def run_user_coro_wrapper() -> None:
                 nonlocal user_result
                 try:
                     user_result = await coro
-                    logger.debug('User coroutine completed successfully.')
                 except Exception as err:
-                    # Log error but don't necessarily stop the server
                     logger.error(f'User coroutine failed: {err}', exc_info=True)
-                    # Store exception? Or let TaskGroup handle it if critical?
-                    # Depending on desired behavior, could raise here to stop everything.
-                    pass  # Continue running server for now
                 finally:
                     user_task_finished_event.set()
 
             reflection_server = _make_reflection_server(self.registry, server_spec)
 
-            # Setup signal handlers for graceful shutdown (parity with JS)
-
-            # Actually, anyio.run handles Ctrl+C (SIGINT) by raising KeyboardInterrupt/CancelledError
-            # For SIGTERM, we might need to be explicit if we run in a container/process manager.
-            # JS uses: process.on('SIGTERM', shutdown); process.on('SIGINT', shutdown);
-
-            # Since anyio/asyncio handles SIGINT well, let's add a task to catch SIGTERM
-            async def handle_sigterm(tg_to_cancel: anyio.abc.TaskGroup) -> None:  # type: ignore[name-defined]
-                with anyio.open_signal_receiver(signal.SIGTERM) as signals:
-                    async for _signum in signals:
-                        logger.info('Received SIGTERM, cancelling tasks...')
-                        tg_to_cancel.cancel_scope.cancel()
-                        return
-
             try:
                 # Use lazy_write=True to prevent race condition where file exists before server is up
                 async with RuntimeManager(server_spec, lazy_write=True) as runtime_manager:
-                    # We use anyio.TaskGroup because it is compatible with
-                    # asyncio's event loop and works with Python 3.10
-                    # (asyncio.TaskGroup was added in 3.11, and we can switch to
-                    # that when we drop support for 3.10).
                     async with anyio.create_task_group() as tg:
                         # Start reflection server in the background.
-                        tg.start_soon(reflection_server.serve, name='genkit-reflection-server')
-                        await logger.ainfo(f'Started Genkit reflection server at {server_spec.url}')
-
-                        # Start SIGTERM handler
-                        tg.start_soon(handle_sigterm, tg, name='genkit-sigterm-handler')
-
-                        # Wait for server to be responsive
-                        # We need to loop and poll the health endpoint or wait for uvicorn to be ready
-                        # Since uvicorn run is blocking (but we are in a task), we can't easily hook into its startup
-                        # unless we use uvicorn's server object directly which we do.
-                        # reflection_server.started is set when uvicorn starts.
-
-                        # Simple polling loop
+                        tg.start_soon(reflection_server.serve)
+                        logger.info(f'Started Genkit reflection server at {server_spec.url}')
 
+                        # Wait for the server to be healthy before starting the user task.
                         max_retries = 20  # 2 seconds total roughly
                         for _i in range(max_retries):
                             try:
-                                # TODO(#4334): Use async http client if available to avoid blocking loop?
-                                # But we are in dev mode, so maybe okay.
-                                # Actually we should use anyio.to_thread to avoid blocking event loop
-                                # or assume standard lib urllib is fast enough for localhost.
-
-                                # Use httpx async client to avoid blocking the event loop
                                 health_url = f'{server_spec.url}/api/__health'
                                 async with httpx.AsyncClient(timeout=0.5) as client:
                                     response = await client.get(health_url)
@@ -207,29 +167,25 @@ async def handle_sigterm(tg_to_cancel: anyio.abc.TaskGroup) -> None:  # type: ig
                         _ = runtime_manager.write_runtime_file()
 
                         # Start the (potentially short-lived) user coroutine wrapper
-                        tg.start_soon(run_user_coro_wrapper, name='genkit-user-coroutine')
-                        await logger.ainfo('Started Genkit user coroutine')
+                        tg.start_soon(run_user_coro_wrapper)
+                        logger.info('Started Genkit user coroutine')
 
                         # Block here until the task group is canceled (e.g. Ctrl+C)
                         # or a task raises an unhandled exception. It should not
                         # exit just because the user coroutine finishes.
-
-            except anyio.get_cancelled_exc_class():
-                logger.info('Development server task group cancelled (e.g., Ctrl+C).')
-                raise
+                        await anyio.Event().wait()
             except Exception:
                 logger.exception('Development server task group error')
                 raise
 
-            # After the TaskGroup finishes (error or cancelation).
+            # After the TaskGroup finishes (normally or by task completion).
             if user_task_finished_event.is_set():
-                await logger.adebug('User coroutine finished before TaskGroup exit.')
-                if user_result is None:
-                    raise RuntimeError('User coroutine finished without a result.')
-                return user_result
+                if user_result is not None:
+                    return user_result
+                else:
+                    raise RuntimeError('User coroutine finished without a result (likely cancelled).')
 
-            await logger.adebug('User coroutine did not finish before TaskGroup exit.')
-            raise RuntimeError('User coroutine did not finish before TaskGroup exit.')
+            return None  # type: ignore[return-value]
 
         return anyio.run(dev_runner)
 
diff --git a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py
index bbcafe7117..999610ebf3 100644
--- a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py
+++ b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py
@@ -161,6 +161,10 @@
     deprecated_enum_metafactory,
 )
 from genkit.plugins.google_genai.models.utils import PartConverter
+from genkit.core.typing import (
+    Candidate,
+    FinishReason,
+)
 from genkit.types import (
     Constrained,
     GenerateRequest,
@@ -1219,7 +1223,7 @@ async def generate(self, request: GenerateRequest, ctx: ActionRunContext) -> Gen
 
         # TODO(#4361): Do not move - this method mutates `request` by extracting system
         # prompts into configuration object
-        request_cfg = self._genkit_to_googleai_cfg(request=request)
+        request_cfg = await self._genkit_to_googleai_cfg(request=request)
 
         # TTS models require response_modalities: ["AUDIO"]
         if is_tts_model(model_name):
@@ -1229,11 +1233,6 @@ async def generate(self, request: GenerateRequest, ctx: ActionRunContext) -> Gen
 
         # Image models require response_modalities: ["TEXT", "IMAGE"]
         if is_image_model(model_name):
-            if request.tools:
-                raise ValueError(
-                    f'Model {model_name} does not support tools. '
-                    'Please remove the tools config or use a model that supports tools.'
-                )
             if not request_cfg:
                 request_cfg = genai_types.GenerateContentConfig()
             request_cfg.response_modalities = ['TEXT', 'IMAGE']
@@ -1397,17 +1396,67 @@ async def _generate(
                 ) from e
             span.set_attribute('genkit:output', dump_json(response))
 
-        content = self._contents_from_response(response)
+        content = await self._contents_from_response(response)
 
         # Ensure we always have at least one content item to avoid UI errors
         if not content:
             content = [TextPart(text='')]
 
+        finish_reason = FinishReason.OTHER
+        candidates = []
+        if response.candidates:
+            for i, c in enumerate(response.candidates):
+                c_content = []
+                if c.content and c.content.parts:
+                    for j, part in enumerate(c.content.parts):
+                        converted = PartConverter.from_gemini(part=part, ref=str(j))
+                        if converted:
+                            c_content.append(converted)
+
+                if not c_content:
+                    c_content = [TextPart(text='')]
+
+                c_finish_reason = FinishReason.OTHER
+                if c.finish_reason:
+                    fr_name = c.finish_reason.name
+                    if fr_name == 'STOP':
+                        c_finish_reason = FinishReason.STOP
+                    elif fr_name == 'MAX_TOKENS':
+                        c_finish_reason = FinishReason.LENGTH
+                    elif fr_name in ['SAFETY', 'RECITATION', 'BLOCKLIST', 'PROHIBITED_CONTENT', 'SPII']:
+                        c_finish_reason = FinishReason.BLOCKED
+                    elif fr_name == 'OTHER':
+                        c_finish_reason = FinishReason.OTHER
+
+                if i == 0:
+                    finish_reason = c_finish_reason
+
+                candidates.append(
+                    Candidate(
+                        index=float(i),
+                        message=Message(role=Role.MODEL, content=c_content),
+                        finish_reason=c_finish_reason,
+                    )
+                )
+
         return GenerateResponse(
             message=Message(
                 content=content,  # type: ignore[arg-type] - content is list[Part] after conversion
                 role=Role.MODEL,
-            )
+            ),
+            finish_reason=finish_reason,
+            candidates=candidates,
+            usage=GenerationUsage(
+                input_tokens=float(response.usage_metadata.prompt_token_count or 0)
+                if response.usage_metadata
+                else None,
+                output_tokens=float(response.usage_metadata.candidates_token_count or 0)
+                if response.usage_metadata
+                else None,
+                total_tokens=float(response.usage_metadata.total_token_count or 0)
+                if response.usage_metadata
+                else None,
+            ),
         )
 
     async def _streaming_generate(
@@ -1520,7 +1569,7 @@ async def _build_messages(
                 continue
             content_parts: list[genai_types.Part] = []
             for p in msg.content:
-                converted = PartConverter.to_gemini(p)
+                converted = await PartConverter.to_gemini(p)
                 if isinstance(converted, list):
                     content_parts.extend(converted)
                 else:
@@ -1540,7 +1589,7 @@ async def _build_messages(
 
         return request_contents, cache
 
-    def _contents_from_response(self, response: genai_types.GenerateContentResponse) -> list:
+    async def _contents_from_response(self, response: genai_types.GenerateContentResponse) -> list:
         """Retrieve contents from google-genai response.
 
         Args:
@@ -1561,15 +1610,21 @@ def _contents_from_response(self, response: genai_types.GenerateContentResponse)
         # Ensure we always return a list, even if empty
         return content if content else []
 
-    def _genkit_to_googleai_cfg(self, request: GenerateRequest) -> genai_types.GenerateContentConfig | None:
-        """Translate GenerationCommonConfig to Google Ai GenerateContentConfig.
+    async def _genkit_to_googleai_cfg(self, request: GenerateRequest) -> genai_types.GenerateContentConfig | None:
+        """Converts a Genkit GenerateRequest to a Gemini GenerateContentConfig."""
+        system_instruction: list[genai.types.Part] = []
 
-        Args:
-            request: Genkit request.
+        # 1. System messages
+        system_messages = list(filter(lambda m: m.role == Role.SYSTEM, request.messages))
+        for m in system_messages:
+            if m.content:
+                for p in m.content:
+                    converted = await PartConverter.to_gemini(p)
+                    if isinstance(converted, list):
+                        system_instruction.extend(converted)
+                    else:
+                        system_instruction.append(converted)
 
-        Returns:
-            Google Ai request config or None.
-        """
         cfg = None
         tools = []
 
@@ -1665,10 +1720,10 @@ def _genkit_to_googleai_cfg(self, request: GenerateRequest) -> genai_types.Gener
 
                 cfg = genai_types.GenerateContentConfig(**dumped_config)
 
-        if request.output:
-            if not cfg:
-                cfg = genai_types.GenerateContentConfig()
+        if not cfg:
+            cfg = genai_types.GenerateContentConfig()
 
+        if request.output:
             response_mime_type = 'application/json' if request.output.format == 'json' and not request.tools else None
             cfg.response_mime_type = response_mime_type
 
@@ -1676,31 +1731,12 @@ def _genkit_to_googleai_cfg(self, request: GenerateRequest) -> genai_types.Gener
                 cfg.response_schema = self._convert_schema_property(request.output.schema)
 
         if request.tools:
-            if not cfg:
-                cfg = genai_types.GenerateContentConfig()
-
             tools.extend(self._get_tools(request))
 
         if tools:
-            if not cfg:
-                cfg = genai_types.GenerateContentConfig()
             cfg.tools = tools
 
-        system_messages = list(filter(lambda m: m.role == Role.SYSTEM, request.messages))
-        if system_messages:
-            system_parts = []
-            if not cfg:
-                cfg = genai.types.GenerateContentConfig()
-
-            for msg in system_messages:
-                for p in msg.content:
-                    converted = PartConverter.to_gemini(p)
-                    if isinstance(converted, list):
-                        system_parts.extend(converted)
-                    else:
-                        system_parts.append(converted)
-            cfg.system_instruction = genai.types.Content(parts=system_parts)
-
+        cfg.system_instruction = system_instruction if system_instruction else None
         return cfg
 
     def _create_usage_stats(self, request: GenerateRequest, response: GenerateResponse) -> GenerationUsage:
diff --git a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/utils.py b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/utils.py
index 304829a323..9381bbb37b 100644
--- a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/utils.py
+++ b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/utils.py
@@ -19,8 +19,10 @@
 import base64
 from typing import cast
 
+import httpx
 from google import genai
 
+from genkit.core.http_client import get_cached_client
 from genkit.core.typing import DocumentPart, Metadata
 from genkit.types import (
     CustomPart,
@@ -62,7 +64,7 @@ class PartConverter:
     DATA = 'data:'
 
     @classmethod
-    def to_gemini(cls, part: Part | DocumentPart) -> genai.types.Part | list[genai.types.Part]:
+    async def to_gemini(cls, part: Part | DocumentPart) -> genai.types.Part | list[genai.types.Part]:
         """Maps a Genkit Part to a Gemini Part.
 
         This method inspects the root type of the Genkit Part and converts it
@@ -164,6 +166,22 @@ def to_gemini(cls, part: Part | DocumentPart) -> genai.types.Part | list[genai.t
                     )
                 )
 
+
+            if url.startswith('http'):
+                try:
+                    data, mime_type = await cls._download_image(url)
+                    # If mime type wasn't in headers, fallback to existing or default
+                    mime_type = mime_type or part.root.media.content_type or 'image/jpeg'
+                    return genai.types.Part(
+                        inline_data=genai.types.Blob(
+                            mime_type=mime_type,
+                            data=data,
+                        )
+                    )
+                except Exception:
+                    # Fallback to file_uri if download fails
+                    pass
+
             return genai.types.Part(
                 file_data=genai.types.FileData(
                     mime_type=part.root.media.content_type,
@@ -240,7 +258,7 @@ def from_gemini(cls, part: genai.types.Part, ref: str | None = None) -> Part:
                         ref=ref or getattr(part.function_call, 'id', None),
                         # restore slashes
                         name=(part.function_call.name or '').replace('__', '/'),
-                        input=part.function_call.args,
+                        input=part.function_call.args if part.function_call.args is not None else {},
                     ),
                     metadata=cls._encode_thought_signature(part.thought_signature),
                 )
@@ -305,3 +323,18 @@ def _encode_thought_signature(cls, thought_signature: bytes | None) -> Metadata
         if thought_signature:
             return Metadata(root={'thoughtSignature': base64.b64encode(thought_signature).decode('utf-8')})
         return None
+
+    @classmethod
+    async def _download_image(cls, url: str) -> tuple[bytes, str | None]:
+        """Downloads an image from a URL.
+
+        Args:
+            url: The URL to download.
+
+        Returns:
+            A tuple containing the image content (bytes) and its MIME type (str or None).
+        """
+        client = get_cached_client(cache_key='google_genai_media')
+        response = await client.get(url, timeout=60.0)
+        response.raise_for_status()
+        return response.content, response.headers.get('content-type')
diff --git a/py/samples/google-genai-hello/src/main.py b/py/samples/google-genai-hello/src/main.py
index 412c9edb3f..1d77798caf 100755
--- a/py/samples/google-genai-hello/src/main.py
+++ b/py/samples/google-genai-hello/src/main.py
@@ -766,8 +766,10 @@ async def tool_calling(input: ToolCallingInput) -> str:
 
 async def main() -> None:
     """Main function - keep alive for Dev UI."""
-    # Keep the process alive for Dev UI
-    _ = await asyncio.Event().wait()
+    await logger.ainfo('Starting main execution loop')
+    while True:
+        await asyncio.sleep(3600)
+    await logger.ainfo('Exiting main execution loop')
 
 
 if __name__ == '__main__':

From a3855cc242a619c09eb773fd91f7479ce1a3e158 Mon Sep 17 00:00:00 2001
From: Mengqin Shen <mengqin@google.com>
Date: Thu, 5 Feb 2026 22:31:50 -0800
Subject: [PATCH 2/3] fix(py): fix lint errors and tox test errors

---
 .../genkit/src/genkit/ai/_base_async.py       | 72 +++++++++++++++----
 .../plugins/google_genai/models/embedder.py   |  6 +-
 .../plugins/google_genai/models/gemini.py     | 23 +++---
 .../plugins/google_genai/models/utils.py      |  4 +-
 .../google-genai/test/google_plugin_test.py   |  5 +-
 py/pyproject.toml                             |  2 +-
 6 files changed, 77 insertions(+), 35 deletions(-)

diff --git a/py/packages/genkit/src/genkit/ai/_base_async.py b/py/packages/genkit/src/genkit/ai/_base_async.py
index c232b57ff8..8f849c65b2 100644
--- a/py/packages/genkit/src/genkit/ai/_base_async.py
+++ b/py/packages/genkit/src/genkit/ai/_base_async.py
@@ -16,8 +16,9 @@
 
 """Asynchronous server gateway interface implementation for Genkit."""
 
+import signal
 from collections.abc import Coroutine
-from typing import Any, TypeVar, cast
+from typing import Any, TypeVar
 
 import anyio
 import httpx
@@ -126,7 +127,7 @@ async def dev_runner() -> T:
             assert spec is not None
             # Capture spec in local var for nested functions (pyrefly doesn't narrow closures)
             server_spec: ServerSpec = spec
-            user_result: T = None  # type: ignore[assignment]
+            user_result: T | None = None
             user_task_finished_event = anyio.Event()
 
             async def run_user_coro_wrapper() -> None:
@@ -134,25 +135,64 @@ async def run_user_coro_wrapper() -> None:
                 nonlocal user_result
                 try:
                     user_result = await coro
+                    logger.debug('User coroutine completed successfully.')
                 except Exception as err:
+                    # Log error but don't necessarily stop the server
                     logger.error(f'User coroutine failed: {err}', exc_info=True)
+                    # Store exception? Or let TaskGroup handle it if critical?
+                    # Depending on desired behavior, could raise here to stop everything.
+                    pass  # Continue running server for now
                 finally:
                     user_task_finished_event.set()
 
             reflection_server = _make_reflection_server(self.registry, server_spec)
 
+            # Setup signal handlers for graceful shutdown (parity with JS)
+
+            # Actually, anyio.run handles Ctrl+C (SIGINT) by raising KeyboardInterrupt/CancelledError
+            # For SIGTERM, we might need to be explicit if we run in a container/process manager.
+            # JS uses: process.on('SIGTERM', shutdown); process.on('SIGINT', shutdown);
+
+            # Since anyio/asyncio handles SIGINT well, let's add a task to catch SIGTERM
+            async def handle_sigterm(tg_to_cancel: anyio.abc.TaskGroup) -> None:  # type: ignore[name-defined]
+                with anyio.open_signal_receiver(signal.SIGTERM) as signals:
+                    async for _signum in signals:
+                        logger.info('Received SIGTERM, cancelling tasks...')
+                        tg_to_cancel.cancel_scope.cancel()
+                        return
+
             try:
                 # Use lazy_write=True to prevent race condition where file exists before server is up
                 async with RuntimeManager(server_spec, lazy_write=True) as runtime_manager:
+                    # We use anyio.TaskGroup because it is compatible with
+                    # asyncio's event loop and works with Python 3.10
+                    # (asyncio.TaskGroup was added in 3.11, and we can switch to
+                    # that when we drop support for 3.10).
                     async with anyio.create_task_group() as tg:
                         # Start reflection server in the background.
-                        tg.start_soon(reflection_server.serve)
-                        logger.info(f'Started Genkit reflection server at {server_spec.url}')
+                        tg.start_soon(reflection_server.serve, name='genkit-reflection-server')
+                        await logger.ainfo(f'Started Genkit reflection server at {server_spec.url}')
+
+                        # Start SIGTERM handler
+                        tg.start_soon(handle_sigterm, tg, name='genkit-sigterm-handler')
+
+                        # Wait for server to be responsive
+                        # We need to loop and poll the health endpoint or wait for uvicorn to be ready
+                        # Since uvicorn run is blocking (but we are in a task), we can't easily hook into its startup
+                        # unless we use uvicorn's server object directly which we do.
+                        # reflection_server.started is set when uvicorn starts.
+
+                        # Simple polling loop
 
-                        # Wait for the server to be healthy before starting the user task.
                         max_retries = 20  # 2 seconds total roughly
                         for _i in range(max_retries):
                             try:
+                                # TODO(#4334): Use async http client if available to avoid blocking loop?
+                                # But we are in dev mode, so maybe okay.
+                                # Actually we should use anyio.to_thread to avoid blocking event loop
+                                # or assume standard lib urllib is fast enough for localhost.
+
+                                # Use httpx async client to avoid blocking the event loop
                                 health_url = f'{server_spec.url}/api/__health'
                                 async with httpx.AsyncClient(timeout=0.5) as client:
                                     response = await client.get(health_url)
@@ -167,25 +207,29 @@ async def run_user_coro_wrapper() -> None:
                         _ = runtime_manager.write_runtime_file()
 
                         # Start the (potentially short-lived) user coroutine wrapper
-                        tg.start_soon(run_user_coro_wrapper)
-                        logger.info('Started Genkit user coroutine')
+                        tg.start_soon(run_user_coro_wrapper, name='genkit-user-coroutine')
+                        await logger.ainfo('Started Genkit user coroutine')
 
                         # Block here until the task group is canceled (e.g. Ctrl+C)
                         # or a task raises an unhandled exception. It should not
                         # exit just because the user coroutine finishes.
-                        await anyio.Event().wait()
+
+            except anyio.get_cancelled_exc_class():
+                logger.info('Development server task group cancelled (e.g., Ctrl+C).')
+                raise
             except Exception:
                 logger.exception('Development server task group error')
                 raise
 
-            # After the TaskGroup finishes (normally or by task completion).
+            # After the TaskGroup finishes (error or cancelation).
             if user_task_finished_event.is_set():
-                if user_result is not None:
-                    return user_result
-                else:
-                    raise RuntimeError('User coroutine finished without a result (likely cancelled).')
+                await logger.adebug('User coroutine finished before TaskGroup exit.')
+                if user_result is None:
+                    raise RuntimeError('User coroutine finished without a result.')
+                return user_result
 
-            return None  # type: ignore[return-value]
+            await logger.adebug('User coroutine did not finish before TaskGroup exit.')
+            raise RuntimeError('User coroutine did not finish before TaskGroup exit.')
 
         return anyio.run(dev_runner)
 
diff --git a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/embedder.py b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/embedder.py
index 32a8a670b7..b5addb7847 100644
--- a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/embedder.py
+++ b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/embedder.py
@@ -94,7 +94,7 @@ async def generate(self, request: EmbedRequest) -> EmbedResponse:
         Returns:
             EmbedResponse
         """
-        contents = self._build_contents(request)
+        contents = await self._build_contents(request)
         config = self._genkit_to_googleai_cfg(request)
         response = await self._client.aio.models.embed_content(
             model=self._version,
@@ -105,7 +105,7 @@ async def generate(self, request: EmbedRequest) -> EmbedResponse:
         embeddings = [Embedding(embedding=em.values or []) for em in (response.embeddings or [])]
         return EmbedResponse(embeddings=embeddings)
 
-    def _build_contents(self, request: EmbedRequest) -> list[genai.types.Content]:
+    async def _build_contents(self, request: EmbedRequest) -> list[genai.types.Content]:
         """Build google-genai request contents from Genkit request.
 
         Args:
@@ -118,7 +118,7 @@ def _build_contents(self, request: EmbedRequest) -> list[genai.types.Content]:
         for doc in request.input:
             content_parts: list[genai.types.Part] = []
             for p in doc.content:
-                converted = PartConverter.to_gemini(p)
+                converted = await PartConverter.to_gemini(p)
                 if isinstance(converted, list):
                     content_parts.extend(converted)
                 else:
diff --git a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py
index 999610ebf3..b48f25c4d9 100644
--- a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py
+++ b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py
@@ -157,14 +157,14 @@
 from genkit.codec import dump_dict, dump_json
 from genkit.core.error import GenkitError, StatusName
 from genkit.core.tracing import tracer
-from genkit.lang.deprecations import (
-    deprecated_enum_metafactory,
-)
-from genkit.plugins.google_genai.models.utils import PartConverter
 from genkit.core.typing import (
     Candidate,
     FinishReason,
 )
+from genkit.lang.deprecations import (
+    deprecated_enum_metafactory,
+)
+from genkit.plugins.google_genai.models.utils import PartConverter
 from genkit.types import (
     Constrained,
     GenerateRequest,
@@ -174,6 +174,7 @@
     GenerationUsage,
     Message,
     ModelInfo,
+    Part,
     Role,
     Stage,
     Supports,
@@ -1400,7 +1401,7 @@ async def _generate(
 
         # Ensure we always have at least one content item to avoid UI errors
         if not content:
-            content = [TextPart(text='')]
+            content = [Part(root=TextPart(text=''))]
 
         finish_reason = FinishReason.OTHER
         candidates = []
@@ -1414,7 +1415,7 @@ async def _generate(
                             c_content.append(converted)
 
                 if not c_content:
-                    c_content = [TextPart(text='')]
+                    c_content = [Part(root=TextPart(text=''))]
 
                 c_finish_reason = FinishReason.OTHER
                 if c.finish_reason:
@@ -1441,7 +1442,7 @@ async def _generate(
 
         return GenerateResponse(
             message=Message(
-                content=content,  # type: ignore[arg-type] - content is list[Part] after conversion
+                content=content,
                 role=Role.MODEL,
             ),
             finish_reason=finish_reason,
@@ -1453,9 +1454,7 @@ async def _generate(
                 output_tokens=float(response.usage_metadata.candidates_token_count or 0)
                 if response.usage_metadata
                 else None,
-                total_tokens=float(response.usage_metadata.total_token_count or 0)
-                if response.usage_metadata
-                else None,
+                total_tokens=float(response.usage_metadata.total_token_count or 0) if response.usage_metadata else None,
             ),
         )
 
@@ -1513,7 +1512,7 @@ async def _streaming_generate(
                 ) from e
         accumulated_content = []
         async for response_chunk in await generator:
-            content = self._contents_from_response(response_chunk)
+            content = await self._contents_from_response(response_chunk)
             if content:  # Only process if we have content
                 accumulated_content.extend(content)
                 ctx.send_chunk(
@@ -1736,7 +1735,7 @@ async def _genkit_to_googleai_cfg(self, request: GenerateRequest) -> genai_types
         if tools:
             cfg.tools = tools
 
-        cfg.system_instruction = system_instruction if system_instruction else None
+        cfg.system_instruction = genai_types.Content(parts=system_instruction) if system_instruction else None
         return cfg
 
     def _create_usage_stats(self, request: GenerateRequest, response: GenerateResponse) -> GenerationUsage:
diff --git a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/utils.py b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/utils.py
index 9381bbb37b..eefa1cb3b4 100644
--- a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/utils.py
+++ b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/utils.py
@@ -19,7 +19,6 @@
 import base64
 from typing import cast
 
-import httpx
 from google import genai
 
 from genkit.core.http_client import get_cached_client
@@ -166,7 +165,6 @@ async def to_gemini(cls, part: Part | DocumentPart) -> genai.types.Part | list[g
                     )
                 )
 
-
             if url.startswith('http'):
                 try:
                     data, mime_type = await cls._download_image(url)
@@ -178,7 +176,7 @@ async def to_gemini(cls, part: Part | DocumentPart) -> genai.types.Part | list[g
                             data=data,
                         )
                     )
-                except Exception:
+                except Exception:  # noqa: S110 - intentionally silent, fallback to file_uri
                     # Fallback to file_uri if download fails
                     pass
 
diff --git a/py/plugins/google-genai/test/google_plugin_test.py b/py/plugins/google-genai/test/google_plugin_test.py
index e6db3feee6..88cd7ad738 100644
--- a/py/plugins/google-genai/test/google_plugin_test.py
+++ b/py/plugins/google-genai/test/google_plugin_test.py
@@ -726,7 +726,8 @@ def test_config_schema_extra_fields() -> None:
     assert config.model_dump()['new_experimental_param'] == 'test'
 
 
-def test_system_prompt_handling() -> None:
+@pytest.mark.asyncio
+async def test_system_prompt_handling() -> None:
     """Test that system prompts are correctly extracted to config."""
     mock_client = MagicMock(spec=genai.Client)
     model = GeminiModel(version='gemini-1.5-flash', client=mock_client)
@@ -739,7 +740,7 @@ def test_system_prompt_handling() -> None:
         config=None,
     )
 
-    cfg = model._genkit_to_googleai_cfg(request)
+    cfg = await model._genkit_to_googleai_cfg(request)
 
     assert cfg is not None
     assert cfg.system_instruction is not None
diff --git a/py/pyproject.toml b/py/pyproject.toml
index 6d718aef8d..cbd6260b0d 100644
--- a/py/pyproject.toml
+++ b/py/pyproject.toml
@@ -75,7 +75,7 @@ dev = [
 lint = [
   "bandit>=1.7.0",
   "deptry>=0.22.0",
-  "litestar>=2.0.0",          # For web/typing.py type resolution
+  "litestar>=2.0.0",     # For web/typing.py type resolution
   "mypy>=1.14.0",
   "pip-audit>=2.7.0",
   "pypdf>=6.6.2",

From 30222744198565caa2e1a748d8346f93bc5db41a Mon Sep 17 00:00:00 2001
From: Mengqin Shen <mengqin@google.com>
Date: Thu, 5 Feb 2026 23:04:51 -0800
Subject: [PATCH 3/3] fix(py): fix nox test errors

---
 .../plugins/google_genai/models/gemini.py     | 88 +++++++++----------
 1 file changed, 40 insertions(+), 48 deletions(-)

diff --git a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py
index b48f25c4d9..cee58628a3 100644
--- a/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py
+++ b/py/plugins/google-genai/src/genkit/plugins/google_genai/models/gemini.py
@@ -1489,7 +1489,7 @@ async def _streaming_generate(
             )
             client = client or self._client
             try:
-                generator = client.aio.models.generate_content_stream(
+                generator = await client.aio.models.generate_content_stream(
                     model=model_name,
                     contents=cast(genai_types.ContentListUnion, request_contents),
                     config=request_cfg,
@@ -1511,7 +1511,7 @@ async def _streaming_generate(
                     cause=e,
                 ) from e
         accumulated_content = []
-        async for response_chunk in await generator:
+        async for response_chunk in generator:
             content = await self._contents_from_response(response_chunk)
             if content:  # Only process if we have content
                 accumulated_content.extend(content)
@@ -1632,13 +1632,11 @@ async def _genkit_to_googleai_cfg(self, request: GenerateRequest) -> genai_types
             if isinstance(request_config, GeminiConfigSchema):
                 cfg = request_config
             elif isinstance(request_config, GenerationCommonConfig):
-                cfg = genai_types.GenerateContentConfig(
-                    max_output_tokens=request_config.max_output_tokens,
-                    top_k=request_config.top_k,
-                    top_p=request_config.top_p,
-                    temperature=request_config.temperature,
-                    stop_sequences=request_config.stop_sequences,
-                )
+                dumped = request_config.model_dump(exclude_none=True)
+                if dumped:
+                    cfg = genai_types.GenerateContentConfig(**dumped)
+                else:
+                    cfg = None
             elif isinstance(request_config, dict):
                 if 'image_config' in request_config:
                     cfg = GeminiImageConfigSchema(**request_config)
@@ -1698,45 +1696,39 @@ async def _genkit_to_googleai_cfg(self, request: GenerateRequest) -> genai_types
                     if key in dumped_config:
                         del dumped_config[key]
 
-                if 'image_config' in dumped_config and isinstance(dumped_config['image_config'], dict):
-                    valid_image_keys = {
-                        'aspect_ratio',
-                        'image_size',
-                        'person_generation',
-                        'output_mime_type',
-                        'output_compression_quality',
-                    }
-                    dumped_config['image_config'] = {
-                        k: v for k, v in dumped_config['image_config'].items() if k in valid_image_keys
-                    }
-
-                # Check if image_config is actually supported by the installed SDK version
-                if (
-                    'image_config' in dumped_config
-                    and 'image_config' not in genai_types.GenerateContentConfig.model_fields
-                ):
-                    del dumped_config['image_config']
-
-                cfg = genai_types.GenerateContentConfig(**dumped_config)
-
-        if not cfg:
-            cfg = genai_types.GenerateContentConfig()
-
-        if request.output:
-            response_mime_type = 'application/json' if request.output.format == 'json' and not request.tools else None
-            cfg.response_mime_type = response_mime_type
-
-            if request.output.schema and request.output.constrained:
-                cfg.response_schema = self._convert_schema_property(request.output.schema)
-
-        if request.tools:
-            tools.extend(self._get_tools(request))
-
-        if tools:
-            cfg.tools = tools
-
-        cfg.system_instruction = genai_types.Content(parts=system_instruction) if system_instruction else None
-        return cfg
+                # Check for SDK support of newer fields
+                for key in ['image_config', 'thinking_config', 'response_modalities']:
+                    if key in dumped_config and key not in genai_types.GenerateContentConfig.model_fields:
+                        del dumped_config[key]
+
+                if dumped_config:
+                    cfg = genai_types.GenerateContentConfig(**dumped_config)
+                else:
+                    cfg = None
+
+        # Tools from top-level field and config-level fields
+        tools.extend(self._get_tools(request))
+
+        if cfg is not None or tools or system_instruction or request.output:
+            if cfg is None:
+                cfg = genai_types.GenerateContentConfig()
+
+            if request.output:
+                response_mime_type = (
+                    'application/json' if request.output.format == 'json' and not request.tools else None
+                )
+                cfg.response_mime_type = response_mime_type
+
+                if request.output.schema and request.output.constrained:
+                    cfg.response_schema = self._convert_schema_property(request.output.schema)
+
+            if tools:
+                cfg.tools = tools
+
+            cfg.system_instruction = genai_types.Content(parts=system_instruction) if system_instruction else None
+            return cfg
+
+        return None
 
     def _create_usage_stats(self, request: GenerateRequest, response: GenerateResponse) -> GenerationUsage:
         """Create usage statistics.